вход по аккаунту



код для вставкиСкачать
Polymer International
Polym Int 48:509±514 (1999)
Thermally stimulated current studies on
molecular relaxation and blow moulding of
poly(ethylene terephthalate)
Rahmat Satoto, Junko Morikawa and Toshimasa Hashimoto*
Department of Organic and Polymeric Materials, Tokyo Institute of Technology, 2-12-1, O-okayama, Meguro-ku, Tokyo 152, Japan
Abstract: The thermally stimulated current (TSC) technique has been used to investigate molecular
relaxation in poly(ethylene terephthalate) (PET) ®lms unstretched and biaxially stretched at 90 and
95 °C. Unstretched PET ®lms show two peaks at 77 and 90 °C corresponding to a and r relaxation
processes, respectively. The a relaxation is associated with the main glass transition of the material.
The r peak with lower intensity is attributable to permanent dipoles. Both biaxially stretched samples
show one TSC peak at 95 °C, supposed to correspond to r relaxation. The disappearance of the a peak,
accompanied by the displacement of the r peak to higher temperature, is the result of the higher
thermal stability of the permanent dipoles, which is strongly in¯uenced by the stiffening of amorphous
parts and the crystallization by stretching. In both stretched samples, the continuous distribution of
pre-exponential factors over activation energies observed might correspond to a single relaxation
mode. The kinematics of stretching PET has been discussed in terms of activation energy and
temperature dependence of relaxation time.
# 1999 Society of Chemical Industry
Keywords: TSC; glass transition; molecular relaxation; relaxation time; activation energy; compensation
The use of poly(ethylene terephthalate) is being
increased because of its versatile nature and competitive price. This material is usually used as containers
such as bottles for cosmetics, medicines, carbonated
and alcoholic beverages and general food (ketchup,
salad dressing, sauce). Stretch blow moulding (SBM)
is widely used to produce such containers. This
method is based on multistage processing. An amorphous injection-moulded preform is reheated to the
rubbery state (between the glass transition temperature Tg, and the onset of thermally induced crystallization) before being stretched biaxially to its ®nal
form. Detailed knowledge of the PET behaviour in the
rubbery state is crucial for the SBM technique.
Rheology has been widely using to investigate
processability of PET though change in molecular
mobility with a very small percentage of additives may
not be detected. Additives (e.g. catalyst, antioxidant,
UV screener, plasticizer) usually exist in the polymer
as a result of polymerization or were added to improve
product performance. The molecular mobility for
bottle processing may vary if the PET resin comes
from different suppliers. Attention is needed to
optimize the production process and to improve the
®nal product performance.
Because of its commercial and technological im-
portance, we have tried to correlate the thermally
stimulated current data with the molecular mobility of
PET, especially in its rubbery state. Even though the
explanations for the TSC results and the relationship
of the activated parameters with the glass transition
phenomena are somewhat controversial,1,2 TSC offers
a capability of studying the effect of additives on
molecular relaxation in low-density poly(ethylene)3
and epoxy resin.4
The thermally stimulated current technique reveals
the mobility of the molecular structure. The rate of
polarization is related to the relaxation time of the
internal motions. This approach provides a new
opportunity to study the physical and morphological
structure of the material, which is important in
studying processability. In this work, global and
thermal sampling TSC techniques were used to
investigate relaxation processes or molecular motion
in poly(ethylene terephthalate). Samples biaxially
stretched at 90 and 95 °C were used as a model of
blow moulding, and unstretched sample was consider
as a preform.
Material used in this work was commercial PET with
an intrinsic viscosity of 0.78 dl g-1, kindly supplied by
* Correspondence to: Toshimasa Hashimoto, Department of Organic and Polymeric Materials, Tokyo Institute of Technology, 2-12-1,
O-okayama, Meguro-ku, Tokyo 152, Japan
(Received 31 July 1998; revised version 26 November 1998; accepted 17 February 1999)
# 1999 Society of Chemical Industry. Polym Int 0959±8103/99/$17.50
R Satoto, J Morikawa, T Hashimoto
Mitsui Petrochemical Co, Japan. Two types of ®lm
were prepared for the measurements, unstretched and
biaxially stretched ®lms. (1) The unstrecthed ®lm
(about 200 mm thick) was prepared by heating the
pellets at 280 °C and extruding them through a ®lm die
to a take-up reel at room temperature. The speed of
the take-up reel was kept the same as that of the
extrudate. (2) For stretched ®lms, the extrudate ®lm
was annealed at (i) 90 °C and (ii) 95 °C for 5 min and
then biaxially stretched at a constant speed of 5 cm s-1
to a draw ratio of 3 3. The thickness of stretched
®lms was about 20 mm.
PET ®lms were cut into 15 mm 15 mm pieces.
Silver paste was applied on both sides of each sample
with a diameter of 8.5 mm, and these were used as
electrodes. Electrical ®elds (Ep) for polarization were
48 kV cm-1 for unstretched samples and 250 kV cm-1
for biaxially stretched samples. Global polarization
was performed by applying a constant electrical ®eld
during cooling at 2 °C min-1 from 100 to 26 °C. In the
case of the thermal sampling (TS) experiments, Ep was
applied at the polarization temperature (Tp) for a time
tp = 5 min. The temperature was then lowered to the
depolarization temperature Td = Tp-5 °C and Ep was
switched off. Depolarization was allowed at a constant
temperature Td for a time td = 5 min. These steps
induced only a certain dipole orientation. The sample
was then quenched to 0 °C to freeze that con®guration.
The electrodes were short-circuited for 6 min. When a
sample was heated at a constant rate of 4 °C min-1, the
TSC was measured using a Keithley 617 electrometer.
The spontaneous TSC was measured by recording the
current during ramp heating of the sample, without
Global TSC curves measured in the range 40±120 °C
with Ep 48 and 250 kV cm-1 for unstretched and
biaxially stretched samples are shown in Fig 1.
Unstretched PET shows two peaks at 77 and 90 °C
corresponding to a and r relaxation processes,
respectively. Both relaxations have been extensively
studied previously,5±10 but the physical signi®cance
towards application was vague. The a relaxation is
associated with the glass transition of the material. The
temperature of the peak increased and its magnitude
decreased with increasing crystallinity.11 This behaviour suggested that the crystals act as crosslinks and
inhibit the movements of the segments in the
amorphous domains. As reported for a semicrystalline
polymer, when the amorphous phase is not at
equilibrium, the glass transition relaxation event can
be related to two peaks Tgl and Tgu (lower and upper
Tg).12,13 The non-equilibrium state can be induced by
drawing the polymer,14 changing the crystallinity of
the polymer15 or by stretched blow moulding.16,17 Tgl
corresponds to the `true' amorphous state of the
polymer. Tgu corresponds to the `constrained' amorphous state. According to the nomenclature used, the
Figure 1. Global TSC spectra of PET films with Tp = 100 –26 °C: curve a,
unstretched, Ep = 48kV cm-1; curve b, biaxially stretched at 90 °C,
Ep = 250 kV cm-1; and curve c, biaxially stretched at 95 °C, Ep = 250kV cm-1.
a relaxation would be associated with Tgl. Based on
volume±temperature curves found by Helwege et al,18
Boyer12 suggested that Tgl is essentially independent of
crystallinity, while Tgu is dependent on it. The thermal
energy required to activate the molecules in the
constrained state is higher, and the resultant second
Tg peak appears at higher temperature. Tg measurements of PET using differential scanning calorimetry
(DSC) gave different values.7,10,19,20 The glass transition of PET was found to linearly decrease with the
increase of crystallinity.20 However, dynamic mechanical analysis (DMA)21 on PET ®lms showed an
increase in Tg with increasing crystallinity. It is
interesting to note the controversial Tg behaviour of
PET. It is certainly open to discussion. The spontaneous TSC peak which appears at 77 °C for unstretched PET (Fig 2) corresponds to the motions
Figure 2. Spontaneous TSC spectra of PET films: curves a and b are of
unstretched films; curve c is for a biaxially stretched film at 95 °C.
Polym Int 48:509±514 (1999)
Molecular relaxation and blow moulding of PET
Figure 3. DSC curves of PET films measured at a heating rate of 200 °C
min-1: curve a, unstretched film; curve b, biaxially stretched film at 90 °C;
and curve c, biaxially stretched film at 95 °C. The arrows from left to right
show the temperatures of glass transition, cold crystallization and melting,
liberated at the glass transition. When the macromolecules relax their `internal stresses', there is motion of
the chains and motion of the dipoles along these
chains. This motion of dipoles in the sample thickness
direction consequently creates a current. The spontaneous TSC peak does not appear in both stretched
The general conclusion is that a and r peaks are
regarded as having different origins of relaxation. The
r peak is regarded as the second Tg, which is
attributable to a kind of dipole. This peak is different
from the r peaks with homopolar character, related to
charge injection observed previously,5,22 because it
appears in the spontaneous TSC (as a negative peak at
around 95 °C). The direction of the peak becomes
positive after polarization (Fig 1), con®rming that it is
not from injection charges.
Biaxially stretched samples show one TSC peak at
95 °C, supposed to be r relaxation. The r peaks shifted
from 90 to 95 °C in both stretched samples. The peak
height was very low for 48 kV cm-1 polarization (not
presented in Fig 1). These phenomena are attributable
to the increasing energy needed to activate the molecular motions. The disappearance of the a peak in
both stretched samples is attributable to the stiffening
of amorphous regions and crystallization by stretching.
Crystallization of the material involves a mobility
restriction of the molecular chains. These data show
good agreement with the measurements by DSC
shown in Fig 3. The shoulder related to the glass
transition temperature in the unstretched sample is
clear, but is not observable in the stretched samples.
The cold crystallization peak appears in unstretched
samples, but is not observed in the stretched samples.
Figure 4 shows a global TSC spectrum (bold line)
and the elementary spectra isolated in unstretched
PET by TS experiments, Tp = 55±90 °C in steps of
Polym Int 48:509±514 (1999)
Figure 4. A global TSC spectrum (bold line) and component TSC spectra
isolated by thermal sampling polarization of unstretched PET films. The
numbers in the figure indicate Tp in °C; tp = td = 5 min, Td = Tp ÿ 5 °C.
Figure 5. Correlation between the temperature of maximum TSC (Tm) and
the polarization temperature (Tp) of PET films: *, unstretched film; &,
biaxially stretched film at 90 °C; and ~, biaxially stretched film at 95 °C.
5 °C. A different specimen was used for each spectrum
and the same Ep of 48 kV cm-1 was applied. All these
elementary spectra can be considered as a single
Debye peak. The relation between the temperature of
maximum TSC Tm, and the polarization temperature
Tp, of each thermal sampling experiments is shown in
Fig 5. The slopes of the straight lines in the Tm versus
Tp are different in both biaxially stretched samples.
However, in both samples, Tm is proportional to Tp,
which is in good agreement with a continuous
distribution in poly(ethylene terephthalate) and
poly(ethylene).23,24 Zielinski and Kryszewski25 proposed theoretical plot of Tm = f(Tp), for the continuous
distributions. Our result in Fig 1 show a broad TSC
maximum, which might be a continuous distribution.
The biaxially stretched sample at 90 °C has a greater
slope than that at 95 °C. The physical meaning of the
R Satoto, J Morikawa, T Hashimoto
value of these slopes is somewhat unclear. The
unstretched sample shows a curve which apparently
consists of two segment lines with different slopes,
indicating that a and r have different origins. These
two slopes may be due to two relaxation modes of
molecular motions.
The temperature evolution of the relaxation times
can well be approached by Arrhenius equation:
…T † ˆ 0 exp…Ea =kT †
where t0 is a pre-exponential factor, k is the Boltzmann
constant, and Ea is the apparent activation energy for
the process. The compensation point is de®ned in the
frequency±temperature relation by two phenomenological parameters: the compensation temperature Tc
and the relaxation time tc.
0 ˆ c exp…ÿEa =kTc †
Figure 7. Plots of pre-exponential factors (to) versus activation energies
(Ea): *, unstretched film, two solid lines; &, biaxially stretched film at
90 °C; and ~, biaxially stretched film at 95°C.
In our calculation Ea was determined by26
Ea ˆ 1:51‰k…Tm T1 †=…Tm ÿ T1 †Š
and the pre-exponential factor22 by
0 ˆ …kTm2 =Ea † exp…ÿEa =kTm †
where T1 is the temperature in which TSC curve has
half the height of maximum TSC, on the lower side of
Tm, and b is the heating rate. The apparent activation
energy Ea calculated from eqn (3) is shown in Fig 6 as a
function of Tp. It is seen that both stretched samples
show the relatively lower activation energy. Sauer and
Kim27 suggested that the lower values of Ea indicate
that the polymer chains are more homogenous in
terms of their structural regularity and/or composition.
This seems to agree with our results in Fig 6 that the
unstretched PET shows the highest Ea and the
stretched PET at 95 °C shows the lowest value. These
three samples have the same composition, but the
structural regularity in the biaxially stretched samples
Figure 6. Variation of activation energy (Ea) as a function of the polarization
temperature (Tp). The arrow indicates the temperature of the a peak:
*, unstretched film; &, biaxially stretched film at 90 °C; and ~, biaxially
stretched film at 95 °C. The unstretched film shows the highest Ea when Tp
is lower than the a peak temperature.
is expected to be much higher than that of the
unstretched sample. Sauer and Kim27 also suggested
that the steepness of the curve ramp is related to the
value of Tg for poly(methyl methacrylate). The PET
stretched at 90 and 95 °C shows substantially the same
ramp, possibly being related to the same Tg for both
samples. As seen in Fig 1, the r peaks appear at the
same temperature.
Figure 7 shows the plot of log t0 versus Ea from the
logarithm of eqn (2). Again the unstretched sample
shows two segment lines (solid lines in this ®gure) and
both stretched PET samples show continuous distributions of pre-exponential factors over activation
energies, indicating that the r peak is composed of a
set of discrete processes close enough to form a single
peak. No resolution of the single processes can be
obtained by means of thermal sampling.25 All the
overlapping processes contribute to the single thermal
sampling maximum. Thus the maximum temperature,
maximum current, natural frequency and activation
energy found from the peak of TS-TSC are not only
determined by a single process, but also by all the
overlapping ones. Figure 7 (compensation line) can be
used to see whether a set of Arrhenius lines obtained at
various Tp values converge to Tc. The good linearity
seen in this plot indicates that the relaxation is
governed by a compensation rule.
The coordinates of the compensation point are
calculated from the slope and intercept of the
compensation line. The result is presented in Table
1. The decrease of tc with orientation suggests that the
size of the moving units decreases.28 The values of
Tc ÿ Tg are almost the same for these three samples,
probably the essence of this parameter is null. The
physical meaning of the compensation temperature is
somewhat controversial.29±37 Lacabanne and coworkers found that Tc ÿ Tg was independent of crystallinity in PET38,39 and poly(ether ether ketone),39
while in other materials such as PET40,41 and
poly(propylene)28 it is related to the degree of crystalPolym Int 48:509±514 (1999)
Molecular relaxation and blow moulding of PET
Table 1. Compensation parameters for PET films
tc /s
Tc / °C
Tc ÿ Tg / °C
Stretched at 90 °C
Stretched at 95 °C
linity of the material. Two compensation points have
been proposed by Ibar,42,43 positive compensation
(Tc > Tg) and negative compensation (Tc < Tg). The
activation energy increases to a maximum when the
temperature increases and approaches Tg; a decrease
of the activation energy with increasing temperature is
expected just above Tg. The negative compensation
point thus seems to be a consequence of this natural
decrease of the activation energy. Moreover, data
pertaining to different methods of analysis could also
describe a compensation law.44 From the diversity of
results found for the glass transition, compensation
and related behaviour in PET, it is certainly open to
discussion. Recently, Sauer and co-workers45,46 emphasized that compensation cannot be related to any
material property for the single reason that it is purely a
result of mathematical manipulation of the Arrhenius
or related equations, and the system is too far underdetermined to allow meaningful extraction of additional parameters such as Tc and tc. We support such a
statement. What can be grasped from the problem is
that dependence of Ea and log to on each other may be
statistical in nature, and probably only one of them
needs to be chosen for analysis.
Figure 8 is a plot representing log t versus 1/T from
the Arrhenius equation. It can be seen clearly after
biaxially stretching that the relaxation time changes to
a longer one (in the rubbery state). At 90 °C, the
relaxation time of unstretched sample is 0.07 s; it 5.64 s
for stretched sample at 90 °c and 12.25 s for stretched
sample at 95 °C. The relaxation times of a PET sample
in the rubbery state are of the same order as the time
for stretch blowing. The change of relaxation times
from 90 to 100 ° (t90/t100) in unstretched samples and
those stretched at 90 and 95 °C are about 58, 10 and 6,
respectively. This means that the molecular mobility in
the unstretched sample is more sensitive to the
temperature change than that in stretched samples.
The larger value of relaxation time in the rubbery state
is correlated with better thermal stability or smaller
molecular mobility. The unstretched sample is supposed to be a bottle preform in blow moulding. When
the sample is being stretched, the effects of changing
temperature and strain rate are highlighted. Bonnebat
et al 47 observed a 15 °C temperature increase in PET
®lm uniaxially stretched at 95 °C within 3 s of
deformation. Warner48 also found similar results in
uniaxially stretched PET. This temperature rise is a
result of plastic deformation from neck formation and
propagation, and crystallization. Concomitant of increasing temperature (from self heating) and crystallization by stretching induced the change in relaxation
time. The relaxation time changes not only when the
sample is stretched, but also after stretching. Maruhashi and Asada49 reported that the best thermal stability
(in terms of shrinkage) in biaxially stretched PET was
found at higher stretching temperature and higher
stretching speed. They proposed that the better
thermal stability is controlled by the larger number
of crystalline regions and the larger relaxed amorphous
regions. We suggest that the higher stretching temperature (associated with higher mobility and shorter
relaxation time) causes molecular segments relax to
and a larger number of crystals to be produced after
cooling. The higher stretching speed facilitates molecular relaxation after stretching, so that the larger
relaxed amorphous form can be expected. The t values
in Fig 8 show that the sample stretched at 95 °C has
slightly better thermal stability than that stretched at
90 °C in the rubbery state.
The TSC technique has been used to analyse molecular mobility in the rubbery state of PET. In the
unstretched (amorphous) sample, the a peak is
associated with the main glass transition of the
material. The r peak with lower intensity is observed
as the second Tg, which is attributable to permanent
dipoles. The molecular mobility in r relaxation is
strongly correlated with the stretching of PET above
Tg. The relaxation time t(T) calculated from the
analysis of the TSC peak shows a relation with stretch
blow moulding. Analysis of dipole relaxation by TSC
is one of the useful techniques for simulating moulding
Figure 8. Arrhenius plot of relaxation time of PET films: *, unstretched
film; &, biaxially stretched film at 90 °C; and ~, biaxially stretched film at
Polym Int 48:509±514 (1999)
1 Sauer BB and Avakian P, Polymer 33:5128 (1992).
R Satoto, J Morikawa, T Hashimoto
2 Lacabanne C, Lamure A, Teyssedre G, Bernes A and Mourgues
M, J Non-Cryst Solids 172:884 (1994).
3 Lei Q, Fan Y, Wang X, Lin J and Xiong Y, J Mater Sci Technol
13:223 (1997).
4 Manggana C and Pissis P, J Macromol Sci-Phys B36:749 (1997).
5 Belana J, Colomer P, Montserrat S and Pujal M, Anal Fis B78:8
6 Belana J, Colomer P, Montserrat S and Pujal M, Anal Fis
B78:142 (1982).
7 Belana J, Colomer P, Pujal M and Montserrat S, J Macromol Sci
Phys B23:476 (1984±85).
8 Belana J, Pujal M, Colomer P and Montserrat S, Polymer 29:1738
9 Belana Punseti J and Colomer Vilanova P, J Mater Sci 26:4823
10 Belana J Mudarra M, Canadas JC and Colomer P, J Mater Sci
28:3805 (1993).
11 Bernes A, Chatain D and Lacabanne C, IEEE International
Symposium on Electrical Insulation, Toronto, Canada 3±6
June 1990, p 457±460.
12 Boyer RF, J Polym Sci Polym Symp 50:189 (1975).
13 Galop M, J Therm Anal 49:107 (1997).
14 Axtell FH and Haworth B, Polym Testing 9:53 (1990).
15 Martin MM, Bernes A and Lacabanne C, J Therm Anal 40:697
16 Axtell FH and Haworth B, Plast Rubber Compos Process Appl
22:127 (1994).
17 DiRaddo RW and Rejon AG, Plast Rubber Compos Process Appl
20:275 (1993).
18 Hellwege KH Hennig Y and Knappe W, Kolloid ZZ Polym 186:29
19 Montserrat S, Colomer P and Belana J, J Mater Chem 2:217
20 Wingard CD, J Therm Anal 49:385 (1997).
21 Foreman J and Reed K, TA Instruments±The TA Hotline 2:2
22 van Turnhout J, Thermally Stimulated Discharge of Polymer
Electrets. Elsevier New York (1975).
23 Hino T, J Appl Phys 46:1956 (1976).
Fischer P and Rohl P, J Polym Sci Phys Ed A14:531 (1975).
Zielinski M and Kryszewski M, Phys Status Solidi a42:305 (1977).
Grosswiener LI, J Appl Phys Rev 24:1306 (1953).
Sauer BB and Kim YH, Macromolecules 30:3323 (1997).
Jarrigeon M, Chabert B, Chatain D, Lacabanne C and Nemoz G,
J Macromol Sci Phys, B17:1 (1980).
Christodoulides C, Apekis L and Pissis P, J Appl Phys 64:1367
Del Val JJ, Alegria A, Colmonero J and Barandiaran JM, Polymer
74:1771 (1986).
Crine JP, J Appl Phys 66:1308 (1989).
Schawb GM, J Catal 84:1 (1983).
Galwey AK, J Catal 84:270 (1983).
Conner WC, J Catal 84:273 (1983).
Exner O, Nature 201:488 (1964).
Garn PD, J Therm Anal 10:99 (1976).
Krug RR, Hunter WG and Greiger RA, J Phys Chem 80:2335
Bernes A, Chatain D and Lacabanne C, Thermochim Acta 204:69
Lacabanne C, Lamura A, Teyssedre G, Bernes A and Mourgues
M, J Non-Cryst Solids 172:884 (1994).
Bernes A, Chatain D and Lacabanne C, Proceedings of the 3rd
International Conference on Properties and Application of
Dielectric Materials, Tokyo, Japan, 8±12, July 1991, p 800±
Bernes A, Chatain D and Lacabanne C, IEEE Trans Electr Insul
Ibar JP, Thermochim Acta 192:91 (1991).
Ibar JP, Polym Eng Sci 31:1467 (1991).
Tesseydre G and Lacabanne C, J Phys D Appl Phys 28:1478
Moura Ramos JJ, Mano JF and Sauer BB, Polymer 38:1081
Sauer BB and Moura Ramos JJ, Polymer 38:4065 (1997).
Bonnebat C, Roullet G and de Vries A J Polym Eng Sci 21:189
Warner SB, J Appl Polym Sci 29:219 (1984).
Maruhashi Y and Asada T, Polym Eng Sci 36:483 (1996).
Polym Int 48:509±514 (1999)
Без категории
Размер файла
104 Кб
Пожаловаться на содержимое документа